We propose a randomized controlled trial (RCT) to test the hypothesis that offering client-centered HIV counseling and testing (HCT) to blood donor candidates will reduce the risk of HIV contamination in the blood supply and also increase appropriate referrals to preventive and care services to persons in need in Sao Paulo, Brazil. Although all donated blood is screened for HIV antibodies, a residual risk of contamination persists due to the immunological window period before antibodies are detectable. Deferral of donors with behavioral risks for HIV is one means to reduce window period contamination; recruitment of voluntary donors from the community (vs. family-replacement donors) is held to be another. Contrary to expectation, a shift to community donors has not resulted in a decrease in HIV prevalence in units of blood collected by our transfusion service. Our preliminary research indicates that some persons at elevated risk use donation as a means of testing for HIV. These test-seeking donors have high trust of the blood bank, do not know other places to test, have low understanding of the immunological window period and believe it is acceptable to deny risk in order to be tested through donation. The test-seeking phenomenon may therefore threaten the safety of the blood supply. Test seeking at blood banks also ill serves persons who need risk reduction counseling because they cannot disclose their true behavior during the donation process and still be tested. Donors also have a low rate of return for test results and therefore do not receive confirmation or referrals to care. To assess whether HCT at the time of donation will improve blood safety and address prevention and care needs, we will randomize donor candidates to be offered this service on-site. First, the impact of the intervention will be measured as an increase in persons receiving their test results, disclosure counseling and referrals (Aim 1). Second, as a biological marker for elevated risk for HIV, we will compare the prevalence of HSV-2 among donors choosing testing to those choosing to donate when offered the choice (Aim 2). Secondary outcomes include differences in prevalence of transfusion-transmitted infections (HIV, HCV, HBV, syphilis, HTLV I/II, Chagas disease), donor motivations (e.g., test-seeking vs. altruism), donor deferral rates, use of confidential unit exclusion, satisfaction with procedures of the blood bank and volume of blood available for use. RCT results will provide rigorous evidence for or against the provision of on-site HCT as an effective means to improve blood safety and link individuals to needed health services. Because our study is integrated into the routine donation system of the largest blood bank in Latin America, results can immediately change donor screening policies and the intervention has feasible scale-up elsewhere in the developing world. Our study assesses both efficacy and effectiveness while helping to strengthen a vital part of Brazil's health system. It is therefore a concrete example of implementation science.